Hi @floswald , currently there are no configuration parameters to enable multiple ssh-keys.
You could think to use a custom bootstrap action to make custom changes to the master and compute instances.
To add to @enrico-usai's answer, you just need to append their public key to the autorized_keys file in your post install script. This is as simple as:
echo "ssh-rsa AAAA...11bh" >> ~/.ssh/authorized_keys
yeah @sean-smith! thanks a ton for that clarification - I was a bit worried that I wouldn't actually know how to attack this issue with a post-install script... now I can certainly achieve that! awesome.
Say, because you are here: my plan was to set up this cluster for a group of colleagues and myself, with different users (and SSH keys) etc. I realized in the tutorial that it foresees only single launches? I.e. you create the cluster, do the work, and you destroy it? (I tried to do cnfcluster stop HelloCluster but that didn't seem to work). So is there no way to 1) stop all compute nodes, 2) stop the master, but 3) preserve the master in "off" state until I need the cluster the next time, just flick the master back on and ready to go?
thanks a lot for your help.
Stopping the cluster only stops the compute nodes. If you'd like to stop the master as well, first stop the cluster:
cfncluster stop HelloCluster
Wait until the compute instances are stopped. You can see the compute instances that are running with:
cfncluster instances HelloCluster
Go to the EC2 Console and find the master instance. You can use the instance id from the previous command. Stop that instance.
To restart the cluster, first start the master instance, wait for the master instance to become available, then run:
cfncluster start HelloCluster
awesome. you made my day sir. that's exactly how I wanted to use this.